Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Autonomous Vehicle Safety Improvement: Less Hype, More Data

687 views

Published on

OESA panel presentation, November 7, 2018

Published in: Technology
  • Be the first to comment

  • Be the first to like this

Autonomous Vehicle Safety Improvement: Less Hype, More Data

  1. 1. Potential Autonomous Vehicle Safety Improvement: Less Hype, More Data OESA Panel, November 7, 2018 © 2018 Philip Koopman Prof. Philip Koopman @PhilKoopman
  2. 2. 2© 2018 Philip Koopman Where did “94%” come from?  “The critical reason was assigned to drivers in an estimated 2,046,000 crashes that comprise 94 percent of the NMVCCS crashes at the national level. However, in none of these cases was the assignment intended to blame the driver for causing the crash.” [DOT HS 812 115] Misquoted: People Cause 94% of Crashes https://www.nhtsa.gov/technology- innovation/automated-vehicles-safety
  3. 3. 3© 2018 Philip Koopman  37461 Fatalities  63% Passenger vehicles  14% Motorcycles  19% Pedestrians, bikes, etc.  3% Large trucks (total 99% due to rounding)  One fatality per 85 million miles, including impaired drivers  28% Alcohol Impairment  28% Seat belt not used in passenger vehicle  27% Speeding  9% Distracted driving  … (total more than 100% due to overlap) Myth: It’s All Cell Phone Distraction (2016 data DOT HS-812-456)
  4. 4. 4© 2018 Philip Koopman  Seat belt use:  Seat belts saved 14,668 lives (age 5+) – 100% seat belt use would save another 2,456 lives  Air bags saved 2,756 (age 13+)  Impaired Driving  Marijuana effects are unclear Can We Improve Seat Belts & Impairment?
  5. 5. 5© 2018 Philip Koopman  AVs can probably cut fatalities in HALF (2014 Casualty Actuarial Society)  22% reduction due to operational limits – Due to: weather, vehicle errors, inoperable traffic control devices  30% reduction due to human behavior – Not wearing seat belts (10% of population; 16% of 8-12 year olds) – Drive manually (in a hurryspeeding) or “I drive better than an AV” – Drive as supervisor while impaired (“it drives itself”) or distracted  AVs are likely to fail differently than people  Social interactions with human drivers & pedestrians is a challenge  Autonomy perception is difficult  AVs have trouble knowing when they don’t know More Realistic AV Safety Expectations https://www.casact.org/pubs/forum/14fforum/CAS%20AVTF_Restated_NMVCCS.pdf
  6. 6. 6© 2018 Philip Koopman  What if full autonomy only deploys in cities?  About 50% of fatalities are in rural areas – About 19% overall rural and not on “arterial” roads  Perhaps 9,500 fatalities avoided for urban improvement of 50%  ADAS as a safety stepping stone  Forward Collision Warning & Automatic Emergency Braking  Driver distraction/inattention monitoring  Need better understanding of semi-autonomy tradeoffs – Supervising semi-autonomy is difficult – Are low level ADAS improvements masking supervision problems? AV vs. ADAS Market Penetration https://www.iihs.org/iihs/topics/t/roadway -and-environment/fatalityfacts/roadway- and-environment/2016

×